Repository navigation
Measure complete Trace constructor parity and decode reduction - #4664
Merged
Merged
Conversation
This was referenced Oct 8, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The native parsing correction in #4537 lacked a complete-constructor comparison against the same full-schema inputs. Add a bounded offline driver comparing literal baseline
0125and reviewed candidatea8c11973on the unchanged preparation05 captured streams. Both versions load their exact three modules under canonical identities; constructor timing is separated from complete state/method checks and decode instrumentation.Actual result: all six constructors returned identical complete typed evidence. Both external payloads retain the same 32,722,082 uncompressed bytes. All six refusal controls, two instance checks and four exact/wrong-command advance checks passed; full-stream decodes fall from two to one per stream. The observed three-sample CPU medians were 0.101295041s baseline and 0.054824542s candidate. Every sample and fixed order are disclosed, including much slower first samples; this is constructor-only evidence, not a generalized speedup, callback fraction or native acceptance result.
The worker completed in 19.36 seconds under its fixed 120-second bound. It starts no native/helper/provider executable, preserves all captured inputs and publishes only to a new disjoint output. Full payloads remain external with compact hashes/timings committed in
notes/native_trace_comparison_2026-10-08/.Validation: 184 comparison/retained-utility tests and 90 production-candidate tests passed. Independent reviews filed and resolved closed-evidence, post-EOF deadline, current-advance coverage and metadata-mutation findings. Final independent inspection verified full payload equality, control graphs, all 68 current case files, 9,214 baseline source files, 83 historical bytes, selected modules/runtime, 17 documented file pins and both JUnit records.
#4537, #4354 and #4400 remain open for separate integration and acceptance. This does not reconstruct unavailable
4366code, relabel original diagnostics or authorize a scientific/native/provider campaign. Candidate PR #4659 has separate CI/integration requirements.Closes #4657
Closes #4658
Closes #4661
Closes #4662
Closes #4663